🌎 Earthquakes: When the Ground Moves

What you’ll learn

1. The Fence That MovedUnderstand elastic rebound — that an earthquake releases energy stored slowly over decades, rather than creating it — and why that makes the loading invisible.After 1906, Harry Fielding Reid noticed that fences near the San Andreas were not just snapped but gently curved for hundreds of feet on either side, and that survey data showed 50 years of slow bending. The earthquake didn't add energy to California; it returned energy California had been storing. The ground is a spring: loading is slow, silent and invisible, release is instant and violent — the asymmetry the whole course hangs from.
2. Locked, Creeping, and the Cycle That Never StopsSee why steady plate motion produces sudden earthquakes — the fault locks, the rock bends, and the debt is always paid.Plates move steadily; faults do not. A locked fault defers motion, and the surrounding rock bends to absorb the difference. Creeping sections like part of the Hayward Fault pay in instalments (offset curbs, a slowly torn stadium); locked sections pay everything at once. Total motion is identical — only the schedule differs. The earthquake cycle then repeats indefinitely, and a rupture's eventual size is decided after it has already begun.
3. What Magnitude Actually MeasuresUnderstand that magnitude measures the source (rigidity × area × slip) and is logarithmic — and why that caps some regions and not others.Moment magnitude comes from seismic moment: how stiff the rock is, how much fault area slipped, and how far it slipped. Because moment scales with area, a ~15 km-deep vertical strike-slip fault caps out near M8 while a shallow-dipping subduction interface can be 1,000 miles by 150 miles — which is why all five of the largest recorded earthquakes are subduction zones. One magnitude unit is ~32× the energy (log E = 1.5M + 4.8), so 'small quakes relieve the strain' would require ~32,000 M4s per M7.
4. The Earthquake Doesn't Hurt You. The Building Does.Separate magnitude from intensity, and understand site effects — amplification, resonance and liquefaction — as where earthquake risk actually lives.Magnitude describes the source and is one number; intensity describes shaking at a place and there are hundreds per earthquake. In 1985 Mexico City, 300+ km from the rupture, was hit harder than towns near it: distance filtered the waves to ~2-second periods, the lake-bed clay resonated at ~2 seconds and amplified them, and mid-rise buildings sway at ~2 seconds too. Liquefaction adds a further failure mode, turning saturated sand briefly into a dense liquid — as in San Francisco's Marina District in 1989.
5. How We Know Where It Was — and What's Inside the EarthUnderstand P and S waves — how their speed difference locates earthquakes, and how the S-wave shadow zone revealed Earth's liquid core.P waves are compressions (~6 km/s, cross anything); S waves are shears (~3.5 km/s) and cannot cross liquid. The S-minus-P gap gives distance, and three stations triangulate an epicentre. The same physics x-rayed the planet: the S-wave shadow zone beyond ~103° proved the outer core is liquid, Mohorovičić found the crust's base in 1909 from waves arriving too early, and Inge Lehmann found the solid inner core in 1936 from faint P waves where none should have been.
6. Prediction Is Not ForecastingUnderstand precisely why earthquake prediction fails and earthquake forecasting works — and that they are different activities, not different precisions.The Parkfield experiment forecast an M5.5–6 before 1993 with 95% confidence and instrumented the fault more densely than any ground in history; the quake came in 2004, and the USGS found no indications that could have been used to predict it. The apparent regularity was a coincidence, and nucleation physics suggests there is no precursor to find. Forecasting survives: UCERF3 gives the SF Bay Area a 72% chance of M6.7+ in 30 years and Los Angeles 60%, which names no date but can be built on.
7. Early Warning Is Neither Prediction Nor MagicUnderstand earthquake early warning as a race won by using a faster medium — and its permanent blind zone.Early-warning systems detect the P wave near the epicentre and send an alert over fibre at ~200,000 km/s while the damaging S wave crawls at 3.5 km/s. This is not prediction: the earthquake has already started. Warning time grows with distance — zero at the epicentre (the blind zone), seconds at 50 km, about a minute at 300 km — which is why Mexico City, threatened by a distant subduction zone, is one of the best-warned cities on Earth.
8. Engineering the Twenty SecondsUnderstand that the hazard is geology but the disaster is construction — and how ductility, soft-storey retrofits and base isolation actually work.Making a building stronger and stiffer is the wrong instinct: unreinforced masonry is strong in compression and nearly helpless in shear, which is why it is the deadliest common building type. Survival comes from ductility — deforming far without losing load capacity. The soft-storey flaw concentrates all demand in an unbraced ground floor (San Francisco's Marina District, 1989), and base isolation inverts the problem entirely by setting the building on horizontally floppy bearings so the ground moves and the building largely doesn't.
9. The GapUnderstand stress transfer and seismic gaps — and be able to state exactly what a gap does and does not license you to claim.A rupture transfers stress to the segments beyond its ends, so earthquakes can queue. Since 1939 the North Anatolian Fault has ruptured in a westward-marching sequence — Erzincan, Niksar, Tosya–Ladik, Bolu, Abant, Mudurnu–Adapazarı, then İzmit in 1999 — leaving an unruptured gap beneath the Sea of Marmara in front of Istanbul that has been loading throughout. That supports a forecast built on bookkeeping (we know the plate rate and the slip owed) but never a prediction: Parkfield's regularity was a coincidence and its quake came 11 years late with no precursor.

Questions this course answers

Grounded in trusted sources

• Harry Fielding Reid, 'The Mechanics of the Earthquake' (Carnegie Institution of Washington, 1910) — elastic rebound theory
• USGS Earthquake Hazards Program — Earthquake Magnitude, Energy Release, and Shaking Intensity
• USGS — 20 Largest Earthquakes in the World (list of Mw values since 1900)
• USGS — M 9.5 1960 Great Chilean (Valdivia) Earthquake event page
• Hanks, T. C. & Kanamori, H., 'A moment magnitude scale', Journal of Geophysical Research (1979)
• USGS — The Modified Mercalli Intensity Scale
• Bakun, W. H. & Lindh, A. G., 'The Parkfield, California, Earthquake Prediction Experiment', Science 229 (1985)
• USGS — The Parkfield, California, Earthquake Experiment; and 'Earthquake prediction lessons from the Parkfield experiment'

Every Wunder lesson is built from real, reputable sources — never invented.

Related History courses